1094(1991)2117-210 © 1991 ElsevierScience PublishersB.V. All rights reserved0167-4889/91/$03.50 016748899100232B
Biochimica et Biophysica Acta,
207
ADONIS
Transforming growth factor beta stimulates the production of the tissue inhibitor of metalloproteinases (TiMP) by human synovial and skin fibroblasts J o h n Kevin W r i g h t , T i m o t h y E d w a r d C a w s t o n a n d B r i a n Leslie H a z l e m a n Rheumatology Research Unit, Unit E6, hddenbrooke "s Hospital, Cambridge ( U.K. ) (Received 4 March 1991) Keywords: TGF-/3;TIMP: Fibroblast;In vitro:Connectivetissuebreakdown IL-I stimulates the secretion of metalloproteinases by a variety of connective tissue cells and is thought to be the primary inducing agent of connective tissue breakdown in rheumatoid arthritis. Transforming growth factor-beta (TGF-IB) is known to be capable of inhibiting the synthesis of metalloproteinases and to be able to partiall~ inhibit interleukin-1 (IL-1) induced cartilage degradation. The present paper examines the ability of TGF.~ to modulate the action of IL-I on fibroblasts of synovial and skin origin and investigates the secretion of the tissue inhibitor of metalloproteiuases (TIMP) by these cells after exposure to TGF./~ and !1.-1. The principal findings are that when four out of frye flbroblast lines were exposed to TGF-/B and IL-I in combination they displayed a significant increase in TIMP secretion; furthermore, in two of these cell lines a significant stimulation of TIMP secretion was induced by TGF-18 alone.
Introduction A family of metailoproteinases including collagenase, stromelysin and gelatinase [1-4] that have been implicated in remodelling of connective tissue are also thought to play a crucial role in the joint destruction seen in rheumatoid arthritis [5,6]. Increased levels of the secreted enzymes or a failure of regulation of their extracellular activity is probably an intrinsic part of the rheumatoid disease process. The control of metalloproteinase activity occurs via a variety of mechanisms including pro-enzyme activation [7] and their inactivation by the specific tissue inhibitor of metalloproteinases (T1MP) [8]. The production of the metalloproteinases can be regulated by polypeptide growth factors and cytokines. The synthesis of procollagenase is known to be affected by agents such as interleukin-g [9], tumour
Abbreviations: DMEM, Dulbecco'smodified Eagle's medium; FCS, foetal calf serum; mRNA, messenger ribonucleic acid; rhlL-l~, recombinanthumaninterleukin-I beta: TGF-~, transforminggrowth factor beta; TIMP, tissue inhibitorof metaUoproteinases. Correspondenco:T.E. Cawston,RheumatologyResearch Unit, Unit E6, Addenbrooke's Hospital, Hills Road, Cambridge, CB2 2OQ, U.K.
necrosis factor [10], platelet derived growth factor [11] and basic fibroblast growth factor [12]. Recently it has been shown that transforming growth factor beta (TGF-/3), a polypeptide produced by a wide variety of cells and tissues [13,14], is capable of inhibiting the synthesis of the metalloproteinases collagenase [12] and stromelysin [15] although, paradoxically, stimulating the secretion of progelatinase [16]. The growth factor is also capable of inhibiting IL-1 induced cartilage breakdown [17] and also of stimulating the production of both TIMP mRNA and of the secreted protein [12,16]. As interleukin-I is considered to play a major role in initiating and maintaining the process of joint destruction in rheumatoid arthritis [9], the present study was carried out to determine whether TGF-fl was capable of modulating the secretion of pro-collagena~:e and TIMP by rheumatoid synoviai fibroblasts stimulated with interleukin-1, as this could have potential therapeutic implications for rheumatoid arthritis and should increase our understanding of the disease process in this disorder. Materials and Methods Cell cul:ure
Synovial tissue, ebtained from patients (age range 24-62 years, female) wire rheumatoid arthritis under-
208 going either therapeutic synovectomy or joint replacement was finely minced and dissociated by treatment with bacterial collagenase (Type CLS III, Worthington) after the method of Dayer et al. [18]. Human skin fibroblasts were obtained from plastic surgery specim,.'ns (age 10, male) using an explant culture method [19]. Both cell types were maintained in vitro in 75 cm 2 flasks (NUNC; GIBCO) in Dulbecco's modification of Eagle's medium (DMEM) (Flow Laboratories) supplemented with 10% foetal calf serum (FCS; Northumbria Biologicals) and antibiotics and antimycotics: gentamicin 50/zg/ml; amphotericin 2.5/~g/ml. The ceils were maintained at saturated humidity in an atmosphere of 95% a i r / 5 % CO2.
Effects of TGF-~ Confluent cells (passage numbers 3-11) were trypsinised and transferred to 24-well culture plates (NUNC, GIBCO) at a density of 5 • 104 cells per well in a total volume of 1 ml of DMEM with 10% FCS. After 24 h the ceils were washed three times with Hanks' balanced salts solution (GIBCO) and serum free DMEM supplemented with 0.01% tissue culture grade bovine serum albumin (Sigma). The cells were maintained in this medium for 48 h at which time the supernatants were removed and fresh serum free DMEM containing either no additions (control), recombinant human interleukin-1/3 (rhlL-1/3, a generous gift of Glaxo Laboratories) alone, and highly purified human TGF-/3 ( R & D Systems, British Biotechnology) in the presence or absence of rhlL-1/3 at the concentrations indicated in the figure legends. All conditions were examined in triplicate. After a further 72 h the culture supernatants were harvested, made 0.02% with respect to sodium azide and stored at -20"C, until assayed as a batch for TIMP and pro-collagenase activity. Protein levels were measured in cell wells at the
1
lO TGF-B
o
o.1
1
lO
ng ml - I
Fig. 2. The secretion of TIMP by HSF8 skin fibronlasts(passage number ~ 3), treated with TGF-/3, in the presence or absence of rhlL-l/3at 1 ng ml- t Statisticaltreatmentwas by the Kruskal-Wallis test with comparisonto the untreated control culture('"P < 0.001).
end of the assay and these were remarkably constant across all treatments indicating that the cell numbers from well to well was not significantly different at the end of the assay period.
Pro.collagenase and TIMP assay Pro-collagenase was assayed using the diffuse fibril method [20] the results being expressed as units of collagenase activity ml-J of supernatant (1 unit = 1 # g of collagen degraded per min at 37*(2). Pro-collagenase was activated by the inclusion of aminophenyl mercuric acetate in the assay at 0.7 raM. TIMP was measured in the diffuse fibril assay by assaying a known amount of active coilagenase in the presence of the culture supernatants (1 unit of TIMP inhibits 2 units of collagenase by 50%).
Statistical analysis The data were analysed using the KruskaI-Wallis test.
Results
e]D-L-IB
.
o.1
6 1 ~ *-.~--'s . . . . . .
TGF-B
TIMP secretion
ng ml - I
Fig. I. The secretionof TIMP by SYI0 synovialfibroblasts(passage number = 9) treated with TGF-/3, in the presence or absence of rhlL-l/~at 1 ng ml- i. Statisticaltreatmentwas by the KruskaI-Wallis test with comparison to the untreated control culture (**P< 0.01, ""P < 0.001).
Four of the five cell lines examined (SYIO, SY28, HSF3 and HSF8) showed a significant stimulation of TIMP secretion when the cells were exposed to a combination of TGF-/3 and rhIL-1/L Furthermore, in the case of two of these cell lines (SY10 and HSF3) a significant stimulation of TIMP secretion was produced by TGF-fl alone. In only one cell line ( S Y l l ) was no effect on TIMP secretion observed in either the presence or absence of rhlL.1/3 (represeatative data shown in Figs. 1-3). With some cell lines (e.g., HSF8) the effects of TGF-/3 in the presence of rhlL-1/3 were dosage dependent, whereas with others (e.g., SY10) the lowest dose of TGF-/3 appeared to induce a maximal response in the levels of TIMP secreted by the cells.
209 !,5 "i
| [] /
.
Discussion
- IL-18
1~"~ +
IL-1B
L
":tJ! iili I0
O
0.1
1
tO
T G F - B ng m)- I
Fig. 3. T h e secretion of T I M P by S Y ) l synovial fibroblasts (passage number=lit treated w i t h TGF-/3 in the presence or absence of r h l L - l / 3 at 1 ng m l - I. No significant differences were found.
No marked differences were observed between synovial fibroblasts and skin flbroblasts in their responses to TGF-/3, although a much larger series of experiments would be required to state this absolutely.
Pro-collagenase secretion With the exception of the HSF8 skin fibroblasts all the cell lines examined displayed a significant increase in pro-collagenase secretion when they were exposed to rhlL-l,8 alone. When the cells were exposed to the combination of TGF-/3 and rhlL-1/3 there was no effect in three of the five cell lines examined on the levels of pro-collagenase secreted by the cells (SYI9, 5Y28 and HSFS). However, in the case of two cell lines (SYI1 and HSF3) there was a significant inhibition of pro-collagenase secretion by TGF-fl (representative data shown in Fig. 4). Again, no differences were observed between the responses of the synovial fibroblasts and ski,:-, flbroblasts to TGF-/3.
+ IL-tB
O.t
I
t0
TGF-B ng ml- I
Fig. 4. The secretion of pro-eollagenase by SYI 1 synovial fibroblasts (passage number = 11), treated with TGF-I,8 in the presence or
absence of rhiL-I/3 at I ng ml-I. Statistical treatment was by the Kruskal-Wallis test with comparisonto the IL-I/3 treatcd/]'GF-~ untreated cultures(+**P < 0.001).
The results demonstrate that TGF-/3 was capable of stimulating the secretion of TIMP by human fibroblasts of synovial or skin origin. Furthermore, this property of the growth factor was most marked when the cells were simultaneously exposed to rhlL-1/3. However, in the case of one cell line (SYll, Figs. 3 and 4) no effect on TIMP secretion was seen either in the presence or absence of IL-I/3. This could indicate that the skin and synovial fibroblasts used in this study require certain cell surface receptors before they can respond and these were absent from some cell lines. Cell-cell interactions or cell-matrix interaetions may be important to prime the cells so that they can respond to cytokines and growth factors. An understanding of the heterogeneity that seems to be present in these cells would give important information concerning the control of TIMP secretion. There is often difficulty in interpreting studies where both enzyme and inhibitor are measured in the same culture supernatant (Ref. 21 and unpublished data). In the present study we observed an increase in "lIMP levels in response to TGF-/L The bioassay for TIMP measures free uncox , xed TIMP. If a simultaneous activation of metalloproteinase occurs, the active enzymes would bind free TIMP reducing the amounts of this protein measured. However, although the significant increases in levels we observed may represent a minimum value there is no doubt that this effect is real and that TGF-/3 increases TIMP production. There was also evidence that, at least in some instances, TGF-/3 was able to inhibit the increases in pro-collagenase secretion seen when these ceils are exposed to IL-1/3. The interpretation of the pro-collagenase results is difficult as the decrease in levels observed could reflect the difficulty of measuring pro-collagenase in the presence of high levels of "lIMP. We are currently developing immunoassays for collagenase and these will show whether this decrease is a true reflection of the amount of enzyme in the culture fluids. Although there have been previous reports of the ability of TGF-fl to stimulate the secretion of TIMP and to inhibit the secretion and synthesis of metalloproteinase enzymes either by itself [16] or in conjunction with other growth factors [12]. Our finding that TGF-/3 can modulate the secretion of these proteins after IL-1/3 stimulation of synovial and skin fibroblasts is apparently a novel one which has important implications for our understanding of the regulation of connective tissue breakdown and our knowledge of the disease process in rheumatoid arthritis. These results shed light on the previous findings of others [23] and those from our own laboratory [17] that TGF-/3 is capable of partially inhibiting the breakdown of porcine articular cartilage stimulated to degrade by
210 IL-1/3. The mechanism of action of IL-l-stimulated cartilage b r e a k d o w n is thought to be m e d i a t e d by metalloproteinases [9,24], although direct evidence of this is still lacking. If TGF-/3 both inhibits metalloproteinase secretion whilst stimulating T I M P secretion by articular chondrocytes this could account for its previously r e p o r t e d ability to inhibit t:artilage d e g r a d a t i o n [17). O u r findings m a y have relevance to the developm e n t of novel therapeutic a g e n t s for the t r e a t m e n t of r h e u m a t o i d arthritis, it might for instance prove feasible to p r o d u c e agents which would mimic the action of TGF-/3 within the inflamed joint. Such agents could serve to ameliorate or even a b r o g r a t e the tissue destruction seen in r h e u m a t o i d arthritis a n d o t h e r disease states in which connective tissue b r e a k d o w n plays a p r o m i n e n t p a r t in the pathology of the the disorder. TGF-/3 is known to p r o m o t e w o u n d healing a n d it is likely that in addition to its k n o w n effects on matrix synthesis an increase in T I M P levels would prevent matrix b r e a k d o w n a n d initiate r e p a i r of connective tissues. Previous strategies for controlling connective tissue b r e a k d o w n have c o n c e n t r a t e d o n diminishing the inf l a m m a t o r y / i m m u n o l o g i c a l c o m p o n e n t s o f such disorders. More recently it has b e e n suggested t h a t the use of metaUoproteinase inhibitors m i g h t prove a useful t h e r a p e u t i c a p p r o a c h . However, these inhibitors would be in the fluid p h a s e a n d may not achieve access to the site of connective tissue destruction. A t h e r a p e u t i c a p p r o a c h aimed at controlling the activities o f connective tissue fibroblasts in a direct fashion, by m e a s u r i n g the local p r o d u c t i o n of T I M P , could prove to be a m o r e a p p r o p r i a t e avenue for f u t u r e endeavours.
Acknowledgements T h e a u t h o r s would like to thank: T h e Arthritis a n d R h e u m a t i s m Council for R e s e a r c h for its g e n e r o u s financial support, Ms. A m a n d a Smith a n d Ms. Louise Julier for their invaluable technical assistance, Mrs. Sheila Smith for her excellent secretarial help in the p r e p a r a t i o n of this m a n u s c r i p t a n d Glaxo Laboratories, G r e e n f o r d , Middlesex for their g e n e r o u s gift of r e c o m b i n a n t h u m a n IL-1B.
References 1 Cawston. T.E.. Mercer, E. and Tyler, J.A. (1981) Bioehira. Biophys. Acta 657, 734-83. 2 Murphy, G., Cawston. T.E.. Galloway, W.A, Barnes. M.J., Bunning, R.A.D., Mercer, E., Reynolds, J.J. and Burgeson, R.E. (1981) Biochem. J. 199, 807-811. 3 Galloway, W.A., Murphy, G., Sandy, .I.D., Gavrilovic, J., Cawston, T.E. and Reynolds, J.J. (1983) Biochem. J. 209, 741-752. 4 Murphy, G., McAlpioe, C.G., Poll, C.T. and Reynolds, J.J. (1985) Biochim. Biophys. Acta 831, 49-58. 5 Harris, E.D. and Krane, 8.M. (1974) New Engl. J. Med. 291, 557-563, 605-609, 652-661. 6 Woolley, D.E., Glanville, R.W., Crossley. M.J. and Evanson, J.M. (1975) Eur. J. Biochem. 54, 611-622. 7 Vater, C.A., Nagase, H. and Harris, E.D. (1986) Biochem. J. 237, 853-858. 8 Cawston, T.E., Galloway, W.A., Mercer, E., Murphy, G. and Reynolds, J.J. (1981) Biochem. J. 195,159-165. 9 Bunoing, R.A.D., Richardson, H.J., Crawford, A., Skjodt, H., Hughes, D., Evans, D.B., Gowen, M., Dobson, P.R.M, Brown, B.L. and Russell, R.G.G. (1986) Agents and Actions (Suppl) 18, 131-152. 10 Dayer, J.M., Beutler, B. and Cerami, A. (1985) J. Exp. Meal. 162. 2163-2168. 11 Bauer, E.A., Cooper, T.W., Huang, J.S., Altman, J. and Deuel, T.F. (1985) Prec. Natl. Acad. Sci. USA 82, 4132-4136. 12 Edwards, D.R., Murphy, G., Reynolds, J.J., Whitham, S.E., Dochert~, A.J.P., Angel, P. and Heath, J.K. (1987) EMBO J. 6, 1899-1904. 13 Massague, J. (1987) Cell 49. 437-438. 14 Sporn, M.B., Roberts, A.B., Wakefield, L.M. and De Crombrugghe, n. (1987) J. Cell Biol. 105,1039-1045. 15 Matrisian, L.M., Lcroy, P., Ruhlmann, C., Gesnel, M.C. and Breathnaeh, R. (1986) Moll. Cell Biol. 6, 1679-1686. 16 Overall, C.M., Wrana, J.L and Sodek,J. (1989) J. Biol. Chem. 264,1860-1869. 17 Andrews, HJ., Edwards, T.A., Cawston, T.E. and Hazleman. B.L. (1989) Biochem. Biophys. Res. Commun. 162,144-150. 18 Dayer, J.M., Krane, S.M., Russell, R.G.G. and Robinson, D.R. (1976) Proe. Natl. Acad. Sci. USA 73, 945-949. 19 Chard, M.D., Wright, LK. and Hazleman, B.L. (1987) Ann. Rheum. Dis. 46, 385-390. 20 Cawston, T.E. and Barrett, AJ. (1979) Anal. Biochem. 99, 340345. 21 Cawston, T.E., Moooey, C.J. and Hazleman, B.L. (1984) Br. J. Rheum. 23, Suppl 136. 22 Reference deleted. 23 Morales, T.I. and Roberts, A.B. (1988) J. Biol. Chem. 263, 12828-12831. 24 Campbell, I.K., Rougbley, P.J. and Mort, J.S. (1986) Bicchem. J. 237, 117-122.
Biochimica et Biophysica Acta,
207
ADONIS
Transforming growth factor beta stimulates the production of the tissue inhibitor of metalloproteinases (TiMP) by human synovial and skin fibroblasts J o h n Kevin W r i g h t , T i m o t h y E d w a r d C a w s t o n a n d B r i a n Leslie H a z l e m a n Rheumatology Research Unit, Unit E6, hddenbrooke "s Hospital, Cambridge ( U.K. ) (Received 4 March 1991) Keywords: TGF-/3;TIMP: Fibroblast;In vitro:Connectivetissuebreakdown IL-I stimulates the secretion of metalloproteinases by a variety of connective tissue cells and is thought to be the primary inducing agent of connective tissue breakdown in rheumatoid arthritis. Transforming growth factor-beta (TGF-IB) is known to be capable of inhibiting the synthesis of metalloproteinases and to be able to partiall~ inhibit interleukin-1 (IL-1) induced cartilage degradation. The present paper examines the ability of TGF.~ to modulate the action of IL-I on fibroblasts of synovial and skin origin and investigates the secretion of the tissue inhibitor of metalloproteiuases (TIMP) by these cells after exposure to TGF./~ and !1.-1. The principal findings are that when four out of frye flbroblast lines were exposed to TGF-/B and IL-I in combination they displayed a significant increase in TIMP secretion; furthermore, in two of these cell lines a significant stimulation of TIMP secretion was induced by TGF-18 alone.
Introduction A family of metailoproteinases including collagenase, stromelysin and gelatinase [1-4] that have been implicated in remodelling of connective tissue are also thought to play a crucial role in the joint destruction seen in rheumatoid arthritis [5,6]. Increased levels of the secreted enzymes or a failure of regulation of their extracellular activity is probably an intrinsic part of the rheumatoid disease process. The control of metalloproteinase activity occurs via a variety of mechanisms including pro-enzyme activation [7] and their inactivation by the specific tissue inhibitor of metalloproteinases (T1MP) [8]. The production of the metalloproteinases can be regulated by polypeptide growth factors and cytokines. The synthesis of procollagenase is known to be affected by agents such as interleukin-g [9], tumour
Abbreviations: DMEM, Dulbecco'smodified Eagle's medium; FCS, foetal calf serum; mRNA, messenger ribonucleic acid; rhlL-l~, recombinanthumaninterleukin-I beta: TGF-~, transforminggrowth factor beta; TIMP, tissue inhibitorof metaUoproteinases. Correspondenco:T.E. Cawston,RheumatologyResearch Unit, Unit E6, Addenbrooke's Hospital, Hills Road, Cambridge, CB2 2OQ, U.K.
necrosis factor [10], platelet derived growth factor [11] and basic fibroblast growth factor [12]. Recently it has been shown that transforming growth factor beta (TGF-/3), a polypeptide produced by a wide variety of cells and tissues [13,14], is capable of inhibiting the synthesis of the metalloproteinases collagenase [12] and stromelysin [15] although, paradoxically, stimulating the secretion of progelatinase [16]. The growth factor is also capable of inhibiting IL-1 induced cartilage breakdown [17] and also of stimulating the production of both TIMP mRNA and of the secreted protein [12,16]. As interleukin-I is considered to play a major role in initiating and maintaining the process of joint destruction in rheumatoid arthritis [9], the present study was carried out to determine whether TGF-fl was capable of modulating the secretion of pro-collagena~:e and TIMP by rheumatoid synoviai fibroblasts stimulated with interleukin-1, as this could have potential therapeutic implications for rheumatoid arthritis and should increase our understanding of the disease process in this disorder. Materials and Methods Cell cul:ure
Synovial tissue, ebtained from patients (age range 24-62 years, female) wire rheumatoid arthritis under-
208 going either therapeutic synovectomy or joint replacement was finely minced and dissociated by treatment with bacterial collagenase (Type CLS III, Worthington) after the method of Dayer et al. [18]. Human skin fibroblasts were obtained from plastic surgery specim,.'ns (age 10, male) using an explant culture method [19]. Both cell types were maintained in vitro in 75 cm 2 flasks (NUNC; GIBCO) in Dulbecco's modification of Eagle's medium (DMEM) (Flow Laboratories) supplemented with 10% foetal calf serum (FCS; Northumbria Biologicals) and antibiotics and antimycotics: gentamicin 50/zg/ml; amphotericin 2.5/~g/ml. The ceils were maintained at saturated humidity in an atmosphere of 95% a i r / 5 % CO2.
Effects of TGF-~ Confluent cells (passage numbers 3-11) were trypsinised and transferred to 24-well culture plates (NUNC, GIBCO) at a density of 5 • 104 cells per well in a total volume of 1 ml of DMEM with 10% FCS. After 24 h the ceils were washed three times with Hanks' balanced salts solution (GIBCO) and serum free DMEM supplemented with 0.01% tissue culture grade bovine serum albumin (Sigma). The cells were maintained in this medium for 48 h at which time the supernatants were removed and fresh serum free DMEM containing either no additions (control), recombinant human interleukin-1/3 (rhlL-1/3, a generous gift of Glaxo Laboratories) alone, and highly purified human TGF-/3 ( R & D Systems, British Biotechnology) in the presence or absence of rhlL-1/3 at the concentrations indicated in the figure legends. All conditions were examined in triplicate. After a further 72 h the culture supernatants were harvested, made 0.02% with respect to sodium azide and stored at -20"C, until assayed as a batch for TIMP and pro-collagenase activity. Protein levels were measured in cell wells at the
1
lO TGF-B
o
o.1
1
lO
ng ml - I
Fig. 2. The secretion of TIMP by HSF8 skin fibronlasts(passage number ~ 3), treated with TGF-/3, in the presence or absence of rhlL-l/3at 1 ng ml- t Statisticaltreatmentwas by the Kruskal-Wallis test with comparisonto the untreated control culture('"P < 0.001).
end of the assay and these were remarkably constant across all treatments indicating that the cell numbers from well to well was not significantly different at the end of the assay period.
Pro.collagenase and TIMP assay Pro-collagenase was assayed using the diffuse fibril method [20] the results being expressed as units of collagenase activity ml-J of supernatant (1 unit = 1 # g of collagen degraded per min at 37*(2). Pro-collagenase was activated by the inclusion of aminophenyl mercuric acetate in the assay at 0.7 raM. TIMP was measured in the diffuse fibril assay by assaying a known amount of active coilagenase in the presence of the culture supernatants (1 unit of TIMP inhibits 2 units of collagenase by 50%).
Statistical analysis The data were analysed using the KruskaI-Wallis test.
Results
e]D-L-IB
.
o.1
6 1 ~ *-.~--'s . . . . . .
TGF-B
TIMP secretion
ng ml - I
Fig. I. The secretionof TIMP by SYI0 synovialfibroblasts(passage number = 9) treated with TGF-/3, in the presence or absence of rhlL-l/~at 1 ng ml- i. Statisticaltreatmentwas by the KruskaI-Wallis test with comparison to the untreated control culture (**P< 0.01, ""P < 0.001).
Four of the five cell lines examined (SYIO, SY28, HSF3 and HSF8) showed a significant stimulation of TIMP secretion when the cells were exposed to a combination of TGF-/3 and rhIL-1/L Furthermore, in the case of two of these cell lines (SY10 and HSF3) a significant stimulation of TIMP secretion was produced by TGF-fl alone. In only one cell line ( S Y l l ) was no effect on TIMP secretion observed in either the presence or absence of rhlL.1/3 (represeatative data shown in Figs. 1-3). With some cell lines (e.g., HSF8) the effects of TGF-/3 in the presence of rhlL-1/3 were dosage dependent, whereas with others (e.g., SY10) the lowest dose of TGF-/3 appeared to induce a maximal response in the levels of TIMP secreted by the cells.
209 !,5 "i
| [] /
.
Discussion
- IL-18
1~"~ +
IL-1B
L
":tJ! iili I0
O
0.1
1
tO
T G F - B ng m)- I
Fig. 3. T h e secretion of T I M P by S Y ) l synovial fibroblasts (passage number=lit treated w i t h TGF-/3 in the presence or absence of r h l L - l / 3 at 1 ng m l - I. No significant differences were found.
No marked differences were observed between synovial fibroblasts and skin flbroblasts in their responses to TGF-/3, although a much larger series of experiments would be required to state this absolutely.
Pro-collagenase secretion With the exception of the HSF8 skin fibroblasts all the cell lines examined displayed a significant increase in pro-collagenase secretion when they were exposed to rhlL-l,8 alone. When the cells were exposed to the combination of TGF-/3 and rhlL-1/3 there was no effect in three of the five cell lines examined on the levels of pro-collagenase secreted by the cells (SYI9, 5Y28 and HSFS). However, in the case of two cell lines (SYI1 and HSF3) there was a significant inhibition of pro-collagenase secretion by TGF-fl (representative data shown in Fig. 4). Again, no differences were observed between the responses of the synovial fibroblasts and ski,:-, flbroblasts to TGF-/3.
+ IL-tB
O.t
I
t0
TGF-B ng ml- I
Fig. 4. The secretion of pro-eollagenase by SYI 1 synovial fibroblasts (passage number = 11), treated with TGF-I,8 in the presence or
absence of rhiL-I/3 at I ng ml-I. Statistical treatment was by the Kruskal-Wallis test with comparisonto the IL-I/3 treatcd/]'GF-~ untreated cultures(+**P < 0.001).
The results demonstrate that TGF-/3 was capable of stimulating the secretion of TIMP by human fibroblasts of synovial or skin origin. Furthermore, this property of the growth factor was most marked when the cells were simultaneously exposed to rhlL-1/3. However, in the case of one cell line (SYll, Figs. 3 and 4) no effect on TIMP secretion was seen either in the presence or absence of IL-I/3. This could indicate that the skin and synovial fibroblasts used in this study require certain cell surface receptors before they can respond and these were absent from some cell lines. Cell-cell interactions or cell-matrix interaetions may be important to prime the cells so that they can respond to cytokines and growth factors. An understanding of the heterogeneity that seems to be present in these cells would give important information concerning the control of TIMP secretion. There is often difficulty in interpreting studies where both enzyme and inhibitor are measured in the same culture supernatant (Ref. 21 and unpublished data). In the present study we observed an increase in "lIMP levels in response to TGF-/L The bioassay for TIMP measures free uncox , xed TIMP. If a simultaneous activation of metalloproteinase occurs, the active enzymes would bind free TIMP reducing the amounts of this protein measured. However, although the significant increases in levels we observed may represent a minimum value there is no doubt that this effect is real and that TGF-/3 increases TIMP production. There was also evidence that, at least in some instances, TGF-/3 was able to inhibit the increases in pro-collagenase secretion seen when these ceils are exposed to IL-1/3. The interpretation of the pro-collagenase results is difficult as the decrease in levels observed could reflect the difficulty of measuring pro-collagenase in the presence of high levels of "lIMP. We are currently developing immunoassays for collagenase and these will show whether this decrease is a true reflection of the amount of enzyme in the culture fluids. Although there have been previous reports of the ability of TGF-fl to stimulate the secretion of TIMP and to inhibit the secretion and synthesis of metalloproteinase enzymes either by itself [16] or in conjunction with other growth factors [12]. Our finding that TGF-/3 can modulate the secretion of these proteins after IL-1/3 stimulation of synovial and skin fibroblasts is apparently a novel one which has important implications for our understanding of the regulation of connective tissue breakdown and our knowledge of the disease process in rheumatoid arthritis. These results shed light on the previous findings of others [23] and those from our own laboratory [17] that TGF-/3 is capable of partially inhibiting the breakdown of porcine articular cartilage stimulated to degrade by
210 IL-1/3. The mechanism of action of IL-l-stimulated cartilage b r e a k d o w n is thought to be m e d i a t e d by metalloproteinases [9,24], although direct evidence of this is still lacking. If TGF-/3 both inhibits metalloproteinase secretion whilst stimulating T I M P secretion by articular chondrocytes this could account for its previously r e p o r t e d ability to inhibit t:artilage d e g r a d a t i o n [17). O u r findings m a y have relevance to the developm e n t of novel therapeutic a g e n t s for the t r e a t m e n t of r h e u m a t o i d arthritis, it might for instance prove feasible to p r o d u c e agents which would mimic the action of TGF-/3 within the inflamed joint. Such agents could serve to ameliorate or even a b r o g r a t e the tissue destruction seen in r h e u m a t o i d arthritis a n d o t h e r disease states in which connective tissue b r e a k d o w n plays a p r o m i n e n t p a r t in the pathology of the the disorder. TGF-/3 is known to p r o m o t e w o u n d healing a n d it is likely that in addition to its k n o w n effects on matrix synthesis an increase in T I M P levels would prevent matrix b r e a k d o w n a n d initiate r e p a i r of connective tissues. Previous strategies for controlling connective tissue b r e a k d o w n have c o n c e n t r a t e d o n diminishing the inf l a m m a t o r y / i m m u n o l o g i c a l c o m p o n e n t s o f such disorders. More recently it has b e e n suggested t h a t the use of metaUoproteinase inhibitors m i g h t prove a useful t h e r a p e u t i c a p p r o a c h . However, these inhibitors would be in the fluid p h a s e a n d may not achieve access to the site of connective tissue destruction. A t h e r a p e u t i c a p p r o a c h aimed at controlling the activities o f connective tissue fibroblasts in a direct fashion, by m e a s u r i n g the local p r o d u c t i o n of T I M P , could prove to be a m o r e a p p r o p r i a t e avenue for f u t u r e endeavours.
Acknowledgements T h e a u t h o r s would like to thank: T h e Arthritis a n d R h e u m a t i s m Council for R e s e a r c h for its g e n e r o u s financial support, Ms. A m a n d a Smith a n d Ms. Louise Julier for their invaluable technical assistance, Mrs. Sheila Smith for her excellent secretarial help in the p r e p a r a t i o n of this m a n u s c r i p t a n d Glaxo Laboratories, G r e e n f o r d , Middlesex for their g e n e r o u s gift of r e c o m b i n a n t h u m a n IL-1B.
References 1 Cawston. T.E.. Mercer, E. and Tyler, J.A. (1981) Bioehira. Biophys. Acta 657, 734-83. 2 Murphy, G., Cawston. T.E.. Galloway, W.A, Barnes. M.J., Bunning, R.A.D., Mercer, E., Reynolds, J.J. and Burgeson, R.E. (1981) Biochem. J. 199, 807-811. 3 Galloway, W.A., Murphy, G., Sandy, .I.D., Gavrilovic, J., Cawston, T.E. and Reynolds, J.J. (1983) Biochem. J. 209, 741-752. 4 Murphy, G., McAlpioe, C.G., Poll, C.T. and Reynolds, J.J. (1985) Biochim. Biophys. Acta 831, 49-58. 5 Harris, E.D. and Krane, 8.M. (1974) New Engl. J. Med. 291, 557-563, 605-609, 652-661. 6 Woolley, D.E., Glanville, R.W., Crossley. M.J. and Evanson, J.M. (1975) Eur. J. Biochem. 54, 611-622. 7 Vater, C.A., Nagase, H. and Harris, E.D. (1986) Biochem. J. 237, 853-858. 8 Cawston, T.E., Galloway, W.A., Mercer, E., Murphy, G. and Reynolds, J.J. (1981) Biochem. J. 195,159-165. 9 Bunoing, R.A.D., Richardson, H.J., Crawford, A., Skjodt, H., Hughes, D., Evans, D.B., Gowen, M., Dobson, P.R.M, Brown, B.L. and Russell, R.G.G. (1986) Agents and Actions (Suppl) 18, 131-152. 10 Dayer, J.M., Beutler, B. and Cerami, A. (1985) J. Exp. Meal. 162. 2163-2168. 11 Bauer, E.A., Cooper, T.W., Huang, J.S., Altman, J. and Deuel, T.F. (1985) Prec. Natl. Acad. Sci. USA 82, 4132-4136. 12 Edwards, D.R., Murphy, G., Reynolds, J.J., Whitham, S.E., Dochert~, A.J.P., Angel, P. and Heath, J.K. (1987) EMBO J. 6, 1899-1904. 13 Massague, J. (1987) Cell 49. 437-438. 14 Sporn, M.B., Roberts, A.B., Wakefield, L.M. and De Crombrugghe, n. (1987) J. Cell Biol. 105,1039-1045. 15 Matrisian, L.M., Lcroy, P., Ruhlmann, C., Gesnel, M.C. and Breathnaeh, R. (1986) Moll. Cell Biol. 6, 1679-1686. 16 Overall, C.M., Wrana, J.L and Sodek,J. (1989) J. Biol. Chem. 264,1860-1869. 17 Andrews, HJ., Edwards, T.A., Cawston, T.E. and Hazleman. B.L. (1989) Biochem. Biophys. Res. Commun. 162,144-150. 18 Dayer, J.M., Krane, S.M., Russell, R.G.G. and Robinson, D.R. (1976) Proe. Natl. Acad. Sci. USA 73, 945-949. 19 Chard, M.D., Wright, LK. and Hazleman, B.L. (1987) Ann. Rheum. Dis. 46, 385-390. 20 Cawston, T.E. and Barrett, AJ. (1979) Anal. Biochem. 99, 340345. 21 Cawston, T.E., Moooey, C.J. and Hazleman, B.L. (1984) Br. J. Rheum. 23, Suppl 136. 22 Reference deleted. 23 Morales, T.I. and Roberts, A.B. (1988) J. Biol. Chem. 263, 12828-12831. 24 Campbell, I.K., Rougbley, P.J. and Mort, J.S. (1986) Bicchem. J. 237, 117-122.
